M24128-BWMN6TP — 128Kbit I²C Serial EEPROM (SOP-8) from STMicroelectronics

What Are the Technical Specifications of M24128-BWMN6TP?

Key numbers that matter: the >4 million write cycles is the spec that separates EEPROM from flash — roughly 40× the endurance of a typical SPI NOR flash. That's the whole reason calibration data lives on an EEPROM. But why does endurance matter more than speed for a parameter store?

The 5ms write cycle is the hidden tax: after every page write the chip goes busy for up to 5ms. Poll for the ACK, don't just fire writes back-to-back.

When Should You Use (and NOT Use) the M24128-BWMN6TP?

✅ Use M24128-BWMN6TP when:

• Frequent, byte-granular data updates. Calibration coefficients, counters, and configuration fields rewritten daily — the 4M-cycle endurance absorbs years of that.
• You're already on an I²C bus. Two wires, shared with sensors and RTCs, no extra SPI CS lines. Address pins A0/A1/A2 let you stack up to eight parts on one bus.
• Data must survive power loss instantly. EEPROM writes are atomic per byte and need no erase step — no block-erase window where a power cut corrupts neighboring data.
• Serial numbers and factory data at 2.5–5.5V. The wide supply range covers 3.3V and 5V logic systems without a regulator on the memory rail.
• Drop-in 24C128 compatibility. The M24128 family follows the standard 24C128 I²C protocol — ST, Microchip, onsemi, and clones share the same address scheme.

❌ Don't use M24128-BWMN6TP when:

• You need more than 16KB. 128Kbit is a parameter store, not a file system. For application data or code, step up to a SPI NOR flash (W25Q32-class).
• You're streaming or logging large blocks. The 64-byte page limit and 5ms write cycle cap throughput at roughly 12KB/s of sustained writes — a SPI flash with 256-byte pages and fast program runs circles around it.
• Your bus is noisy or long. I²C tops out around 400kHz–1MHz and needs pull-ups; a noisy harness wants SPI or a differential bus instead.
• Writes are rare but you need huge capacity. If the data changes once at the factory and never again, a cheaper SPI flash (100K-cycle endurance) is plenty.

What Are the Alternatives to M24128-BWMN6TP?

The 24C128 vs SPI flash decision in one line: EEPROM for data that changes often and must survive power cuts; SPI flash for data that's mostly written once and read many times. The 4M-cycle endurance gap is the deciding number.

What Are the Typical Applications of M24128-BWMN6TP?

Calibration and parameter storage: sensor offsets, gain coefficients, and trim values written at the factory and updated in the field. The 4M-cycle endurance absorbs firmware updates that rewrite the block a hundred times a day.

Serial numbers and device identity: one page holds the MAC address, serial, and manufacture date, written once and read at every boot. The 200-year retention means the identity outlives the product.

Configuration memory on shared I²C buses: alongside an RTC, a temperature sensor, and a battery monitor, the EEPROM rides the same two wires. The A0/A1/A2 pins let multiple EEPROMs coexist at different addresses.

Industrial and medical data logging: counters, event flags, and last-known-good state written on every power transition. EEPROM's atomic byte writes survive mid-write power loss without corrupting the neighboring data — something flash's block-erase can't promise.

Why Buy M24128-BWMN6TP from ICMASS?

Every lot tested for address decode and write behavior. We batch-verify the device responds at 0x50 and completes a full page write before shipping. Counterfeit and re-marked 24C128s commonly fail page-write wrap or read back garbage at the wrong address.

Cross-reference support for the 24C128 family. Not sure whether the board wants ST, Microchip, or onsemi? Send us your bus speed and capacity requirement — we'll tell you which part drops into the footprint.

BOM consolidation for control boards. The same product usually carries an MCU, an RTC, and the M24128. One shipment from Shenzhen covers the whole digital section.

Same-day dispatch, 5–10 days worldwide. Orders before 15:00 CST ship same day via DHL or FedEx. For volume orders, we source directly from the ST production line.

Frequently Asked Questions About M24128-BWMN6TP

Q1: My MCU can't find the EEPROM on the I²C bus — what's wrong?

A: Run an I²C scanner and check the address pins — most "missing" EEPROMs are at a different address than the code expects. With A0/A1/A2 grounded the part answers at 0x50; any high pin shifts it (0x52, 0x54, etc.). Address pins must be tied to VCC or GND — floating inputs behave unpredictably. Missing 4.7k pull-ups cause the same symptom.

Q2: Writes are silently rejected — why?

A: The WC pin is high, or the pull-up is missing. Tie pin 7 to GND to enable writes; tied high, the whole array is hardware-write-protected and every write is silently ignored. This is the first hardware check in every EEPROM troubleshooting thread.

Q3: My code works with a 24C02 but hangs with the M24128 — what changed?

A: The 128Kbit part needs a 2-byte internal address. Small EEPROMs like the 24C02 address within one byte; the 24C128/24C256 family sends two address bytes after the device address. Firmware that skips the second byte stalls on the first random-access operation.

Q4: What pull-up resistor value do I need?

A: 4.7kΩ on both SDA and SCL — the internal MCU pull-ups aren't enough. At 400kHz–1MHz the bus capacitance charges through the pull-ups; too weak and edges round off, too strong and the bus draws excess current. 4.7k is the standard starting point for 3.3V and 5V buses.

Q5: My page write wraps around and corrupts data — what's happening?

A: You're crossing the 64-byte page boundary in a single write. Page writes wrap within the page — bytes past the boundary land at the page start. Split writes that span pages into separate page-write operations; this is the classic 24C128 page-write bug reported on forums.

Q6: How long do I wait between writes?

A: Up to 5ms — poll the ACK instead of delaying blindly. After a write the chip goes busy and won't ACK new start conditions. The robust pattern is ACK-polling: keep sending the device address until it ACKs, then continue. A fixed 5ms delay also works but wastes time on every write.

Q7: Can I put several EEPROMs on one bus?

A: Yes — up to eight, by setting A0/A1/A2 differently on each. Give each part a unique address and the bus supports them all. On longer buses watch the total capacitance: roughly one device per 1m of wire at 400kHz, or drop the clock rate.

Q8: Can a 3.3V MCU talk to a 5V bus EEPROM?

A: Yes — the M24128 runs from 2.5V to 5.5V, so power it at the MCU's rail. If the bus has 5V devices pulling up to 5V, put the EEPROM on the same 5V rail and level-shift the MCU side, or use series resistors. Don't pull a 3.3V-powered EEPROM's lines up to 5V.

Q9: Is 4 million write cycles really enough?

A: For parameter storage, comfortably — do the math. A device rewriting 10 bytes every minute does 14,400 writes a day: 4M cycles lasts 277 days if hammering one address, but with wear leveling across 16KB it's effectively never worn out. Flash's 100K cycles would fail the same pattern in a week of concentrated writes.

Q10: M24128 vs a SPI flash — which should I use?

A: Frequent small writes → EEPROM; big data or code → SPI flash. The EEPROM wins on endurance (4M vs 100K cycles), byte-granular writes with no erase step, and two-wire integration. The flash wins on capacity (32Mbit vs 128Kbit) and raw write throughput. See our EEPROM vs SPI Flash comparison for the full decision guide.